Figure 3

Illustration of the clusters of archaeological samples from the cemetery and the settlement in the plane of the PCA main components for the selected feature subset 1 (a) and subset 2 (b). Each ellipse is a 95% confidence ellipse for a 2D normally distributed data set. The probability of samples belonging to clusters 1 or 2 based on the selected acid data is also presented for feature subset 1 (c) and subset 2 (d). Confusion matrices for feature subset 1 (e) and subset 2 (f).